“Having been one of the lucky few who lived through the disaster, I could see rescue authorities going around trying their best to rescue people,” Wani said. “However, the scale of the disaster was so large that they couldn't identify who needed to be rescued first, and what the priorities for rescue were.”
First responders were burdened by a lack of proper tools to coordinate efforts and clear blocked roadways to extricate victims. The aftermath’s resulting pandemonium — and subsequent national and international earthquakes that followed — led Wani, along with his fellow Stanford University alumni Nicole Hu and Timothy Frank, to create One Concern. The startup aspires to be one of the first to use artificial intelligence to save lives through analytical disaster assessment and calculated damage estimates. Wani said that with the platform, emergency operations centers (EOCs) can receive instant recommendations on response priorities and other insights to dispatch resources effectively.
One Concern’s efforts to pioneer machine learning services for state and local agencies has earned it a spot in Government Technology's GovTech100, a list of noteworthy companies to watch in the public-sector IT market.
Wani, who serves as One Concern’s CEO, elaborated on his startup’s origins and how the company is progressing after its recent beta launch.
Government Technology: What led to the idea and ultimate conception of One Concern?
One Concern CEO Ahmad Wani: The combination of the right people, the right background and a unifying problem. I am from Kashmir, a region prone to earthquakes and floods. When I was 17 years old, in 2005, 70,000 people lost their lives in an earthquake in my hometown. This event compelled me to study engineering and specifically in 2005, start performing earthquake engineering research.
Then, in 2014, a combination of two events on different sides of the world inspired the creation of One Concern.
In 2014, during a break from graduate school at Stanford, I was visiting my parents in Kashmir when a large flood engulfed the state. Eighty percent of Kashmir went under water within minutes! Most people were on their rooftops for up to seven days without food and water, waiting for rescue. Thousands of people lost their lives ... most of the rescue was random, and response priorities were ad hoc. I was quick to assume that the problem of situational awareness was probably restricted to the developing world.
Thereafter, upon my return to the Bay Area, I discovered that a natural disaster had struck in California, in Napa County, while I was away. A magnitude 6.0 earthquake on Aug. 24, 2014, in Napa had driven thousands of 911 calls, and overwhelmed the authorities who carried out the rescue effort on a first-come, first-served basis. Power and communication lines were not fully restored for weeks. In fact, most of the badly damaged regions did not even have connections to a phone network and couldn’t call in for help — which has been seen in nearly all of the previous earthquakes.
Napa was just a magnitude 6.0 earthquake, which caused minor or no damage in most regions. I seemed to imagine the situation in a magnitude 7.0 earthquake (nearly 10 times larger than a magnitude 6.0) or a magnitude 8.0 (nearly 100 times larger), for which there is significant probability of occurrence in California. Not just Napa, but 25 other counties could be very badly hit and out of network. I recognized at that point there would be a huge need for situational awareness, for driving rescue and relief operations.
It was clear that difficulties in the chaos following a natural disaster were not limited to the developing world. It is a worldwide issue and it affects all regions vulnerable to any kind of natural disaster. Not only rescue, but the importance of reconnaissance and recovery also became apparent.
These two events crystallized for me the opportunity to solve the problem of post-disaster reconnaissance and rescue through artificial intelligence given the community scale and highly nonlinear nature of the problem. First up was earthquakes, as I was studying earthquake engineering at Stanford. I partnered with Nicole Hu and Timothy Frank, who were both driven to solve this problem based on their unique previous experiences. Nicole, at the time a computer science grad student, had worked for large companies in data science and Web security, but she was looking for a real-world problem that could really make a positive impact on people's lives. Timothy, a major in the U.S. Air Force and a Ph.D. candidate in structural engineering, had several years of experience in emergency management and disaster response. He was well aware of the current tools and processes in addition to the challenges and pain points that those in the emergency response community face. Like Nicole and me, Timothy was also driven by the impact this project could have on saving lives and strengthening communities.
A combined class project between the Machine Learning and Performance Based Earthquake Engineering courses at Stanford was undertaken. Results of the project were compared with historical records. The goal was to see if damage to homes could be predicted on a community scale for the Napa earthquake. Results showed that our machine learning model accurately predicted structural damage to homes on a scale of 1-to-4 with significant accuracy.
At the public presentation for the machine learning course, a venture capitalist — who turned out to be our first investor and trusted adviser and mentor — Mar Hershenson of Pejman Mar Ventures, stopped by to chat. She said if we were interested in starting a company, to let her know. Then, Stanford Machine Learning Professor Andrew Ng, famous for his pioneering roles with the Google Brain project and Coursera, stopped by to see our project. He immediately saw a need for it in the real world, and offered his support and advice to start a company, and to develop the class project into something more robust. It didn't take the three of us long to commit to taking the project forward and see if we can make a difference to communities all over California, the U.S. and the world.
Government Technology: How did One Concern select the data sources that are analyzed to create damage estimates post-disaster?
Wani: Our team of skilled and experienced domain experts in the fields of earthquake engineering and flood modeling have carefully selected our data sources. We are becoming a repository for organized geospatial data in the process.
Government Technology: What actionable intelligence does One Concern provide its users and how does it do it?
Wani: Our vision is to change the way society plans for, responds to and recovers from all types of natural disasters. Currently we provide critical situational awareness in the minutes and hours following an earthquake. Our core product is a Web platform called “Seismic Concern” that alerts you when an earthquake may have affected your jurisdiction, and displays a color-coded map of the likely structural damage. This saves time in reconnaissance and allows emergency operations centers to allocate their limited resources to rescue and recovery. It gets all the stakeholders a common operating picture and not only facilitates response prioritization, but also recovery operations such as material staging and shelter management. Apart from the map, the platform furnishes key insights like the elderly population in a particular block that is badly damaged, or the number of kids in a school which could be hit. This helps in instant situational awareness and assigning response priorities to the places that need the most help. For instance, an elementary school wouldn’t be prioritized if an earthquake struck at midnight, but if it hit at 10 a.m. on a schoolday, it might jump to the top of the response priority list.
Moreover, going to the phone network being jammed or lines being down, most people who have the worst damage cannot call 911. Seismic Concern is built on geographically distributed servers and redundant systems, which means our site will be up during an emergency when the emergency operations center needs it most.
Compiling an Initial Damage Estimate (IDE) is critical for emergency operation centers to request financial aid from state level and federal authorities. Initially done using a windshield tour, wherein emergency managers carry out a quick reconnaissance survey of the area, street by street, it took around two months of time after the [2014] Napa earthquake to complete the process of requesting financial aid. "Seismic Concern" would provide a scientific basis to an IDE, an incident commander would be able to identify and quantify extent of damage to his jurisdiction with a significant amount of accuracy in minutes, thus saving a lot of time, and promising high precision.
We use state-of-the-art machine learning algorithms and stochastic modeling on derived features, and proprietary models to do this for the people we serve, which as of now are exclusively local governments.
Apart from the live response platform, One Concern also offers a training module, which works on the same artificial intelligence and stochastic modeling back end of Seismic Concern combined with state of art geophysical and seismological research. This technological breakthrough will enable emergency operations centers to train on scenarios based on actual simulations to get a real sense of the situation. The platform would provide damage from simulated earthquakes for emergency response drills before a disaster strikes. This can aid in personnel readiness and plans development, thereby making a community more resilient.
We view the Seismic Concern Training module as democratizing the information in our platform, and allowing local agencies to inform, train and empower the public how to respond to predictive disaster scenarios. The training module is a powerful tool for public policy makers and emergency operation centers in their interactions with the public at large.
Government Technology: How is One Concern working to develop its platform through its recent beta launch?
Wani: We are receiving feedback from our beta partners that will allow us to improve the user experience, as well as add functionality to support the needs of the community. What we provide should change the way emergency managers do business. It should be easy to use, simple to understand, and facilitate action that will save time, money and lives. We won't stop until that mission is achieved.
Government Technology: What user input did you gather from cities and first responders to create the analytic and predictive analytics tools
Wani: We interviewed dozens of experts in the emergency management community from all over the U.S. and beyond. We learned the foremost problem immediately following a disaster is situational awareness. All stakeholders need rapid, accurate information about the situation in a format easy to see and use. I'm sure we'll be making modifications and additions to our platform, but from some initial feedback, we hit the nail on the head with the foremost problem. We learned about how the emergency operations centers operate, how first responders and 911 call centers do their heroic jobs, how a diverse team of professionals work behind the scenes to coordinate response and recovery, how levels of emergency managers (e.g., city, county, state, federal) relate and report to each other to share resources and receive aid. The business of being ready and able to respond to anything at any time is a huge industry that spans the globe. The people we are working with are an invaluable resource to help us develop our platform. We expect our product to be one of their most valuable resources in their time of crisis. We empower heroes with the actionable information they need to save lives.
Government Technology: How would you describe One Concern's current development as a startup? For example, are you in the seed stage, reaching out to angel investors, or preparing for a round of Series A venture capital?
Wani: We raised a seed round in the fall of 2015, grew our team, and deployed and refined the product. Based on feedback from our partners, we then built our training module. We'd like to reach out to any city or county who may be interested in being an early adopter of this revolutionary technology. We can be contacted at aboutus@oneconcern.com.
